Anisotropy in pore networks analyzed with 2-D autocorrelation (variomaps)

This paper describes a system for generating variomaps (autocorrelation or variance as a topological map) of binary images. The images are produced from plane sections through porous media. The 2-D autocorrelation function is calculated by Fourier transformation. The resulting map of the autocorrelation illustrates the mean porel size and directionality (anisotropy) in the analyzed section. The method has been tested on porespace combined within simulated grain packs, two types of sandstones, and catalyst material. The software is written in Pascal for PC systems and circumvents the DOS 640K barrier, thus enabling the handling of even large images. Calculation time for a 512 × 512 Fourier transform is less than 2 min on a 33 MHz 486 PC.